You are on page 1of 15

exp

10
12
-2

exp
7.8
12

100

100

30
Step 1
Step 2
Evaluation Metrics Step 3(a)
Strep 3(b)
Step 3 (c)

Yes
No

Pass
Fail

Stu 1
Stu 100

KNN
Dt
X1
2
10
sala
10000
12000
-2000

sala
10000
12000

Tra

Testng
Ensemble - Combinig existing algorithms -
1. Bagging
2. Boosting
3. Stacking

Dataset - Visualiuzation; Descriptive stats


Preprocess
Regression - Supervised - linear, polynomial
Classification - Supervised - KNN, DT, "ENSEMBLE"
Clustering - Unsupervised - partioning - k-means, k-modes; Hierarchical - Agglomerative/divisive

Positive - more importatnt


Negative

Predict
Yes
No
No
Yes

9
7
X2 (Y)
mar m f y
2 1 0 2.2
2 0 1 3
4 1 0 4

male male 1
female female 0
male

min max
2 100

X1 X2 Y-predict Y
12 10 2
13 10 3

Only RMSE

100 70 R2=0.9
Validation/test - 1 Mock Test 30 RMSE= 1.75

Test Data - 2 Real Exam Test RMSE


y-actual - y_predicted 5 0.7
y-actual - y_average 10

1-0.7 0.3
Step 4:
MAE, MSE, RMSE, R2
Confusion Matrix
Silhoutte

Yes/No
Yes-Positive
No- Negative

Actual - y Predict
Row 100 Yes Yes TP
Row 101 Yes No FN
Row 102 No No TN
Row 103 No Yes FP
Row 104 No No TN

Accuracy 1+2/5 0.6

Precision 1/1+1 0.5

Actual
No FN
Yes FP Critical As less as possible
No TN
Yes TP
X3 (Z) X4
age salary
0 0.1 0.9
1 0.2 0.92
0.001 0.3

5.78

0 1

/RMSE R2
2.5 0.3 8.0% of variance in y values are correctly predicted by the linear regression
5.5

LR

RMSE=1.5
he linear regression

You might also like